Deep Dual-resolution Networks for Real-time and Accurate Semantic Segmentation of Road Scenes

Introduction

This is the unofficial code of Deep Dual-resolution Networks for Real-time and Accurate Semantic Segmentation of Road Scenes. which achieve state-of-the-art trade-off between accuracy and speed on cityscapes and camvid, without using inference acceleration and extra data!on single 2080Ti GPU, DDRNet-23-slim yields 77.4% mIoU at 109 FPS on Cityscapes test set and 74.4% mIoU at 230 FPS on CamVid test set.

The code mainly borrows from HRNet-Semantic-Segmentation OCR and the official repository, thanks for their work.

requirements

Here I list the software and hardware used in my experiment

pytorch==1.7.0
3080*2
cuda==11.1

Quick start

0. Data preparation

You need to download the Cityscapesdatasets. and rename the folder cityscapes, then put the data under data folder.

└── data
  ├── cityscapes
  └── list

1. Pretrained model

download the pretrained model on imagenet or the segmentation model from the official，and put the files in ${PROJECT}/pretrained_models folder

VAL

use the official pretrained model and our eval.py code. with ydhongHIT's advice now can reach the same accuracy in the paper. Thanks.

cd ${PROJECT}
python tools/eval.py --cfg experiments/cityscapes/ddrnet23_slim.yaml

model	Train Set	Test Set	OHEM	Multi-scale	Flip	mIoU	Link
DDRNet23_slim	unknown	eval	Yes	No	No	77.83	official
DDRNet23_slim	unknown	eval	Yes	No	Yes	78.42	official
DDRNet23	unknown	eval	Yes	No	No	79.51	official
DDRNet23	unknown	eval	Yes	No	Yes	79.98	official

Note

with the ALIGN_CORNERS: false in ***.yaml will reach higher accuracy.

TRAIN

download the imagenet pretrained model, and then train the model with 2 nvidia-3080

cd ${PROJECT}
python -m torch.distributed.launch --nproc_per_node=2 tools/train.py --cfg experiments/cityscapes/ddrnet23_slim.yaml

the own trained model coming soon

OWN model

model	Train Set	Test Set	OHEM	Multi-scale	Flip	mIoU	Link
DDRNet23_slim	train	eval	Yes	No	Yes	77.77	Baidu/password:it2s
DDRNet23_slim	train	eval	Yes	Yes	Yes	79.57	Baidu/password:it2s
DDRNet23	train	eval	Yes	No	Yes	~	None
DDRNet39	train	eval	Yes	No	Yes	~	None

Note

set the ALIGN_CORNERS: true in ***.yaml, because i use the default setting in HRNet-Semantic-Segmentation OCR.
Multi-scale with scales: 0.5,0.75,1.0,1.25,1.5,1.75. it runs too slow.
from ydhongHIT, can change the align_corners=True with better performance, the default option is False